On Thu, 19 Sep 2002, Kevin P. Fleming wrote:
> I've got callout verification caching code writting and (nearly)
> working. As it turned out, it was quite easy to do, owing to how well
> organized and understandable the Exim sources are. (I'm working with
> Exim 4.10, of course).
Excellent. It is always (well, usually :-) less work for me to
incorporate a patch than to do something like this from scratch.
> After that, it attempts to write additional records when needed, but
> they never show up in the database file. dbfn_write is returning a
> result of 13, but I can't figure what that's supposed to mean.
What DBM library are you using? I'm not really an expert on any of them,
but I would guess that this might depend on which one it is.
> If you're out there Philip, give me a pointer as to what you think might
> be going on. When this issue is resolved, I'll post a patch to the list
> for others to try out.
I'm here today, but away most of next week. If it isn't resolved by
then, you'll have to send me the patch to try, but at the moment I'm
giving 99.9% of my attention to getting the book produced (estimated to
take the next month).
But I'm just about to go to the dentist. I'll take a quick look at the
source when I subsequently get in to work.
> The feature list is:
> - all callout verification results in records being added to the hint
> database, unless the callout defers
> - records are marked as positive or negative
> - two new config options, "callout_positive_expire" (default 24 hours)
> and "callout_negative_expire" (default 2 hours) control how long records
> are kept in the database
That's exactly the spec I had in mind.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
